Meta has invested heavily in artificial intelligence and the data centers needed to run it. Bloomberg reported on Wednesday that Meta is developing plans to create a cloud infrastructure business that would sell both AI compute capacity and access to hosted models.
What the plan would do and why it matters
According to Bloomberg, the move would position Meta as a cloud provider competing with established players such as Amazon Web Services, Google Cloud and Microsoft Azure. The idea is to convert Meta’s large, capital-intensive data-center investments into a more immediate revenue stream.
Industry context and similar deals
Bloomberg notes that Meta’s reported plans come weeks after similar capacity-leasing activity from SpaceX. In early May, SpaceX and xAI signed a deal with Anthropic to buy out all compute capacity at SpaceX’s Colossus 1 data center, and SpaceX has since made comparable leases with Google and Reflection AI.
Those developments suggest that control of physical data-center capacity could be a decisive advantage in the AI market, not only who builds the best models or services.
Risks and skepticism
Some industry observers warn that rapid AI infrastructure build-outs risk creating a bubble reliant on chips that depreciate quickly. Others question whether AI companies can generate enough end-user revenue to justify the multibillion- or trillion-dollar bets being placed on infrastructure and model development.
Meta’s scale of investment and current position
Bloomberg reports that as of the end of the first quarter Meta had committed $182.9 billion to AI infrastructure spending over coming years. That total includes large ongoing projects in Louisiana and Ohio. The Ohio facility — which Mark Zuckerberg likened to the size of Manhattan — is expected to come online this year.
Unlike companies such as Google and OpenAI, Meta has not yet seen substantial external demand for its own AI models and services. Meta does not break out revenue from Meta AI or the Llama open-weight model family in its earnings reports, and company executives have generally emphasized internal, corporate uses of AI. That suggests Meta’s AI efforts may not yet represent a material standalone revenue line.
How Meta might monetize its capacity
Bloomberg says Meta may emulate CoreWeave by selling access to raw compute capacity. The outlet also reports Meta is considering following AWS’s example by selling access to various AI models hosted on its infrastructure, including its recently launched closed-weight model, Muse Spark.
The new business line is reportedly being developed under an initiative called Meta Compute and would be led by Santosh Janardhan, head of infrastructure; Daniel Gross, head of Meta Superintelligence Labs; and Dina Powell McCormick, Meta’s president.
Next steps and confirmation
Bloomberg’s reporting echoes Mark Zuckerberg’s May comments that a Meta cloud computing business is “definitely on the table” as a way to recoup some of the company’s massive AI investments. TechCrunch has reached out to Meta for comment.
In short, Meta appears to be exploring turning its excess data-center capacity into a commercial cloud offering that would sell compute and hosted models, a move that would reuse its costly infrastructure investments while testing whether data-center ownership can be a durable competitive advantage in AI.
